Core Browser Agent

0•

Automates web browser interactions and page analysis to enable Claude to browse websites and interact with web elements directly.

The core-agent-browser skill provides Claude with comprehensive browser automation capabilities via the agent-browser CLI. It allows the AI to navigate URLs, inspect page structures through interactive accessibility tree snapshots, and perform complex interactions like clicking, form filling, and scrolling using unique reference IDs. While often used as a fallback for specialized research agents, it is a powerful standalone tool for situations where Claude needs to dynamically explore, test, or extract data from any web interface without pre-defined selectors.

주요 기능

01Interactive DOM snapshots with reference IDs for precise element targeting
02Advanced wait conditions for network idle states, text presence, or element loading
030 GitHub stars
04Comprehensive interaction suite for clicking, typing, hovering, and scrolling
05Full browser navigation control including back, forward, reload, and close
06Visual verification through standard and full-page screenshots

사용 사례

01Performing interactive end-to-end testing and visual debugging of web interfaces
02Automating multi-step form submissions and complex web authentication flows
03Extracting documentation or data from websites lacking structured APIs or pre-computed selectors